Replace every square-bracket placeholder with the public company, source or date in scope.
Private-company research questions for human and AI agents.
A private-company research agent should begin with public facts, dated source links, freshness checks and explicit gaps. These prompts help investors, advisers, corporate-development teams and agent builders structure first-pass research while keeping private diligence, recommendations, matching, ranking, valuation and suitability out of scope. Every material fact still needs source checks and human review.
Information only. Not investment advice. Not an offer of securities in any jurisdiction.
Set the evidence rules before asking the question.
Set an as-of date so freshness and conflicts can be judged against a clear point in time.
Require a dated source for every material fact and leave unsupported fields explicitly unknown.
Keep public research separate from authorised private diligence and require human review.
Paste this before any prompt in the pack.
Use only public, source-linked information available for [COMPANY_NAME] as of [AS_OF_DATE]. Separate verified facts from interpretation. For every material fact, provide the source URL and source or access date. Mark stale, conflicting or missing information explicitly. Do not infer revenue, funding, valuation, customers, ownership, investor interest or suitability. Do not recommend, rank, score or match companies. End with facts that need human verification before any private diligence request.
Choose the research job, then copy a question.
Understand the company
Start with the public description and separate each claim into a sourceable field.
01 Describe what [COMPANY_NAME] publicly says it does in one sentence. Then separate its customer problem, product or service, sector and geography into distinct fields. Cite each claim.
02 Summarise the public profile at [COMPANY_PROFILE_URL] as of [AS_OF_DATE]. Return verified facts, source links, source dates and unresolved gaps.
03 Which public statements support the description of [COMPANY_NAME] as operating in [SECTOR]? Use only claims supported by cited sources.
Check evidence and freshness
Make dates, conflicts and unknown fields visible before relying on a profile.
01 List the dated public sources supporting [COMPANY_NAME]'s profile. Group them as current, potentially stale or conflicting relative to [AS_OF_DATE]. Explain the date rule used.
02 For each material profile field for [COMPANY_NAME], return the latest supporting [SOURCE_DATE], [SOURCE_URL] and a freshness note. Mark any field without evidence as unknown.
03 Identify contradictions between the public profile at [COMPANY_PROFILE_URL] and cited public sources. Do not resolve a contradiction by guessing. State what a human should verify.
Review public traction signals without ranking
Describe observable evidence without turning it into an investment conclusion.
01 Summarise dated public traction signals for [COMPANY_NAME], such as product releases, hiring pages, public partnerships or published case studies. Do not infer revenue, customer concentration, valuation or investment quality.
02 Compare the types and recency of public evidence for [COMPARISON_COMPANY_A] and [COMPARISON_COMPANY_B]. Do not rank, score or recommend either company.
03 Which public claims about [COMPANY_NAME] are supported by primary sources, which rely on secondary sources and which remain unsupported?
Find gaps before private diligence
Keep missing public information explicit and leave restricted evidence to an authorised process.
01 From public sources only, list material facts that remain unknown for [COMPANY_NAME]. Classify each as missing, stale or conflicting. Do not invent the answer.
02 Create a human-verification checklist from gaps in public evidence for [COMPANY_NAME]. Exclude requests for cap tables, customer names, personal contact details, credentials or private documents from this public workflow.
03 State which questions about [COMPANY_NAME] cannot be answered from the public profile and should be deferred to an authorised private diligence process.
Prepare a first-pass research note
Use a consistent structure that keeps facts, interpretation and follow-up separate.
01 Prepare a first-pass note on [COMPANY_NAME] with public description, sector, geography, dated evidence, observable traction signals, conflicts, missing facts and human follow-up. Keep facts and interpretation separate.
02 Turn the cited public evidence for [COMPANY_NAME] into a one-page briefing with an as-of date of [AS_OF_DATE] and a source log. Exclude recommendations, ranking, suitability, valuation and matching.
03 Summarise what is known publicly about [COMPANY_NAME] and what needs authorised verification before deeper diligence. Do not treat [MISSING_FACT] as a negative signal.
Run a repeatable MCP or API workflow
These are natural-language patterns, not current endpoint, field or response-schema claims.
01 Using only fields currently exposed through authenticated public-profile access, retrieve [PUBLIC_PROFILE_ID] as of [AS_OF_DATE]. Return available facts, source links, freshness dates and missing fields.
02 Search for public profiles in [SECTOR] and [GEOGRAPHY]. Return only fields the authorised tool exposes, explain filters and avoid ranking or recommendation language.
03 Before using an MCP tool or API for [COMPANY_NAME], list the requested public fields and allowed scope. After retrieval, cite returned sources, flag missing fields and require human review for conflicts.
04 Repeat the same public-profile question for [COMPARISON_COMPANY_A] and [COMPARISON_COMPANY_B]. Keep output consistent and compare evidence coverage only, not investment quality.
Test public-profile readiness for founders
Identify which public facts need clearer structure, sources or freshness dates.
01 Review [COMPANY_PROFILE_URL] for clarity, source coverage and freshness. Identify missing public facts that would help a human or agent understand the company.
02 Which public-profile claims for [COMPANY_NAME] need a dated source link? Do not request private documents, customer identities, personal contact details, cap-table data or credentials.
03 Create a public-profile update checklist for [COMPANY_NAME] covering description, sector, geography, source links, freshness dates and explicit unknowns.
Keep facts, sources, gaps and review status separate.
Content example only. This is not an API or MCP contract and does not describe a current HubbleField response schema.
{
"company": "[COMPANY_NAME]",
"as_of": "[AS_OF_DATE]",
"facts": [{
"claim": "[PUBLIC_FACT]",
"source_url": "[SOURCE_URL]",
"source_date": "[SOURCE_DATE]"
}],
"gaps": ["[MISSING_FACT]"],
"review_required": true,
"boundary": "Public information only"
}Keep public discovery permission-aware.
Public research workflow
- Use structured public company profiles, dated public sources and explicit unknowns.
- Keep every retrieved field within the current authenticated public-profile scope.
- Treat HubbleField as an information-only discovery surface that still requires source checks.
Authorised private diligence
- Do not expose private documents, cap tables, customer names or investor identities.
- Do not include personal contact details, credentials, unpublished profiles or restricted records.
- Move questions that need non-public evidence into a separate authorised private diligence process.
Check the evidence before using the note.
- Set an as-of date.
- Cite each material fact.
- Distinguish primary and secondary sources where relevant.
- Label stale, conflicting and missing information.
- Never fill gaps with model inference.
- Confirm retrieved fields are within the authenticated scope.
- Keep public discovery separate from authorised private diligence.
- Require human review of material facts and conflicts.
- Exclude recommendation, ranking, matching, suitability and valuation conclusions.
Move from public questions to repeatable retrieval.
HubbleField is an information-only discovery surface for structured public company profiles. Authenticated API and MCP access can make approved public-profile research more repeatable, but the output still requires source checks and human review.
What should a private-company research agent ask first?
Start with what the company publicly says it does, which dated sources support that description, how fresh each material fact is and which facts remain unknown or conflicting.
Can these prompts be used with MCP or an API?
Yes, when an authorised tool exposes the relevant public-profile fields. Confirm the current tool, field and access scope, then require source links, freshness dates and human review.
Are these prompts a substitute for private diligence?
No. They structure first-pass research from public information. Restricted information belongs in a separate authorised diligence process.
Does HubbleField recommend or rank companies?
No. HubbleField provides information-only discovery. It does not provide investment advice, recommendations, ranking, matching, suitability assessment or valuation.
What should an agent do when information is missing?
Mark it unknown, stale or conflicting, cite available evidence and state what a human should verify. Do not invent or infer the missing fact.
Choose the path that matches your role.
Information only. Not investment advice. Not an offer of securities in any jurisdiction.